Image Tools
Two optional plugins bring images into a conversation, and each has its own reference page.
Choose the plugin
| Plugin | Tool | Use it to |
|---|---|---|
image-gen |
GenerateImage |
Create a new image from a text description. |
image-edit |
EditImage |
Change an existing PNG or JPEG from the current Project or a public URL. |
- Create a picture from a description covers the
image-genplugin and itsGenerateImagetool. - Image Editing covers the
image-editplugin and itsEditImagetool.
Install either plugin from Settings → Available plugins, and check its marketplace entry for the core release it needs. Both need an image provider configured in the AI settings section, which is named after your assistant, and selected per plugin under Settings → Plugins.
The image model chooses the size, and each result uses the format the provider returns: PNG, JPEG or WebP.
How the two plugins work
Both plugins do one thing with two tools: they name the image provider you selected and ask Chetty to render the picture. Chetty holds the provider's key or account token and writes the finished file into the Project, so a plugin never sees a credential.
Each call makes one image request and saves one file. There is no retry and no polling, and the result is not attached to the conversation — showing a saved image is a separate, explicit ShareImage step.
An output path and overwrite are checked before the provider is called. A matching extension is kept, a mismatched one is corrected to the format the provider actually returned, and the tool reports the path it wrote.
Note: If this installation has no configured provider, the plugin registers no tool, so Chetty is not offered it.